Here is what I discovered after wondering if there’s a different or better alternative to chatgpt.
I have been aggressively using chatgpt for quick and grounded decision making, especially on financial matters and business related decisions, including on whether to quit a new job, start a new one, or just leverage what comes naturally. The results have been incredible so far. But what is chatgpt generations without field tests? I have gone ahead to run the results in real time, and i can say chat is fuckin impressive.
So it got me wondering, am i sidelining other options? I looked up alternatives, and this is what I deduced in two days.
I took Perplexity and Claude. I tried hitting the three (chat included) with a business idea with a long term maturity.. This is what I learned.
Perplexity is the most stupid or useless ai tool of the list. Just regurgitation and quoting sources, end of story.
On the other hand, Claude clocked in second. But wait a minute. It almost took first position hadn’t i noticed one thing. This bastard is too positive on its outlook of things. And that Is its undoing. Unless you are the type of person who loves vibes and just wants rhetorical assurances and absconding consequence-bearing responsibilities, you will be impressed by this LLM. It does cite sources and try its best to think and do analyses, but it’s positiveness does get in the way of, wait for it! Reasoning! It does reason, but not in a realistic way. That right there, is the drawback.
Now let’s go to the most famous beast. Chatgpt. By now I know you are familiar with this bad boy. The one who is too quick to fool. Try some unfounded prompting and see if you can fool it. It has plenty of drawbacks, however. The main reason for me wanting other tools was chat’s retardedness of trying to link everything it knows about me as an advantage to the task at hand. Something retarded and unrelated I’d say. It’s like a Claude-type-micro-moment glitch but at the end. Too much unnecessary reliance on positive outlook at past info it has of our discussions. It made me question if chat is stupid and i need another perspective. Enter the underperforming competition.
The other gripe i noticed stemmed from chatgpt’s behavior was constant reliance on its general knowledge to try and use it as a basis for making a case of its argument. This bitch confidently claims a stance with general knowledge, but the good thing is that such taught me that to counter this, everything must be countered with a request for citations. The answers improved significantly and offered better scopes. Including pulling some really hidden sources.
On the other hand, it doesn’t sugarcoat things. The reasoning is super grounded and realistic minus the faux positivity of Claude. The business at hand had a timeline of 2 years on chat, 6 months on Claude. Just like every question they both handled in the same sentence structure all the way down to spelling. All the way down to a control scenario where i reviewed a different business via similar wording on the two platforms
So I’m back to chat but with refined scopes. What do you think of LLMs, or chat in short?